saltymimir

6 mesi fadiscuss

The obvious culprit here are the LLMs, but I do wonder whether Github's social features, despite its flaws, have given developers fewer reasons to ask questions on SO?

Speaking from experience, every time I hit a wall with my projects, I would instinctively visit the project's repo first, and check on the issues / discussions page. More often than not, I was able to find someone with an adjacent problem and get close enough to a solution just by looking at the resolution. If it all failed, I would fall back to asking questions on the discussion forum first before even considering to visit SO.

saltymimir

2 anni fadiscuss

Finally finish a workable infinite canvas + decentralized knowledge graph project. Been way too distracted with other subplots of the project the past few months.

As for the non-technical stuffs, I'd like to at least breach the 25-minute mark in my 5ks (just under 26 minutes flat is fine). Running's been such a revelation for me this year. Definitely agreed with the idea that avoiding burnout is mostly a matter of doing the inverse of what you do on a regular basis[1].

[1]: https://gwern.net/backstop#burnout

saltymimir

2 anni fadiscuss

Pleasantly surprised that this site’s still around. I remember seeing this back in college, and it was one of the catalysts that got me into building sites and this career in the first place.

(I’m 30 btw, so this site’s quite old by internet standards.)

saltymimir

2 anni fadiscuss

These are the slides that are about to be presented (or has been presented already?) by some of the TC39 committee members.

Some genuinely interesting bits about how the committee sees the entire Javascript ecosystem today and what they envision the future will (probably) look like.

saltymimir

3 anni fadiscuss

Ohh boy, as someone who works on a travel booking, service serving customers in Asia, this problem hits hard. I guess the answer depends on how you plan to handle the names once they got submitted.

I happen to be handing that data over to airlines, which has some of the less forgiving, yet fragmented name requirements. If you handle this incorrectly, your customer can't fly, even after they paid for the flight. And for those who say that this doesn't matter as much: It absolutely does. People do get confused by this more frequently than you think. I've seen people losing an entire trip that they saved for, all because of unclear naming requirements.

The way I deal with this is to provide a country and locale specific name fields. You don't have to detect the geolocation or track the user for this, just let them choose whatever locale setting they want, and give them the "sensible" layout. Here are some examples:

- In Vietnam, we use last name then first name.

- In Indonesia, we use first name, then last name, but also give an option to declare that the person doesn't have a last name.

- In Singapore, we use a single field to input the first name and last name.

Even when you've handled the layout convention carefully, the 3rd party you're handing the data to, if one exists, might not give the same care and attention that you do. In my case: some airlines just haven't gotten around the idea that some people simply don't have last names. When a person with a single name wants to fly, airlines want the customer to use the name for both first and last name (e.g. If the person's name is David, then the airline expects "David David"). If you require First Name and Last Name as the input, and don't elaborate on how to fill them, the customer might simply fill the last name with a dot (".") character. The airlines / any other 3rd party won't accept that. For this, I suggest to detail out the ways in which you handle the data and go talk to your providers, if any.

All in all, it's a pretty tough challenge, and the wisdom around this isn't going to fit inside a single HN post. I do commend you for actually thinking about this problem. Good Luck.

saltymimir

3 anni fadiscuss

> As with previous Macedonian snake cults, the focus of worship at the temple was on fertility. Barren women would bring offerings to Glycon in hopes of becoming pregnant. According to Lucian, Alexander had less magical ways of causing pregnancy among his flock as well.

Wondering what they are :)

saltymimir

3 anni fadiscuss

Here's a bookmarklet I've been using to submit HTML page data to a webhook endpoint: https://gist.github.com/theAJFM/c15535c0b7a994b126f41e127466...

It extracts the page's title, the url, and even prompts you to give it some tags if you need some.

You still need to host your own webhook to save it, but that too should be straightforward. Just use firefox's readability.js to strip the bloat elements and save it in your database. You could save it into Airtable like what OP's been doing, or even save it as a file in your cloud storage.

saltymimir

3 anni fadiscuss

I decided to take a look at the redesigned mobile app after some time not using it. Can't believe how much worse it got.

I don't appreciate their move towards "social" feed sure, but the reader has genuinely gotten worse. I remember that it used to have the options to adjust:

- Line Height

- Page Width

- Text Alignment

- Fonts beyond the defaults (for paid subscribers, which I am one).

In the new app, it was all gone. Yet, I'm still paying the same premium price for it. The web app still has those perks, but it's probably because there hasn't been any update to it for the past 2 years or so.

I haven't had any use for the app for some years now. Like what many say in this thread, it's not that hard to build a system that replaces it well enough. You just need to make it scrape links, turn them to markdown (personal preference), and index the entire content for FTS in SQLite.

Thanks to the OP for reminding me to cancel my subscription. Now I just need to find a way to export my huge pocket archive before the subscription renews.

saltymimir

3 anni fadiscuss

The magazine has ceased publication, but I really enjoyed reading https://reallifemag.com/.

saltymimir

4 anni fadiscuss

Was a bit skeptical about how good this was gonna be, but I'm definitely impressed.

The improvements are a lot more apparent on desktop. I love the fact that I can do pretty much anything using the command palette and keyboard shortcuts. Feels like this is the kind of browsing experience that I'm most contend with. The GPT-3 "ghostreader" feature was also great; most of the summary / text generations fulfilled my expectations.

If I have to pick on something: the mobile app browsing experience isn't that much better from Pocket or Instapaper. The scrolling and animation feels a bit laggy in my iPhone. The "ghostreader" feature in the app feels very limited and awkward to enable here as well.

saltymimir

4 anni fadiscuss

In my (admittedly) brief experience dabbling with asciidoc, I really don't like the opinionated approach on how styles should be injected to the generated HTML output.

Say that I generate a table using asciidoc's built-in syntax, the output will always generate HTML elements with the CSS classes already defined:

    <table class="tableblock frame-all grid-all stretch">
      <colgroup>
        <col style="width: 50%;">
        <col style="width: 50%;">
      </colgroup>
      <thead>
        <tr>
          <th class="tableblock halign-left valign-top">Column heading 1</th>
          <th class="tableblock halign-left valign-top">Column heading 2</th>
        </tr>
      </thead>
      <tbody>
        <tr>
          <td class="tableblock halign-left valign-top"><p class="tableblock">Column 1, row 1</p></td>
         <td class="tableblock halign-left valign-top"><p class="tableblock">Column 2, row 1</p></td>
        </tr>
        <tr>
          <td class="tableblock halign-left valign-top"><p class="tableblock">Column 1, row 2</p></td>
          <td class="tableblock halign-left valign-top"><p class="tableblock">Column 2, row 2</p></td>
        </tr>
      </tbody>
    </table>

For me, this output is needlessly verbose for a setup that uses nothing but default flags. Worse, it doesn't provide much of a leeway to change the classes or even let me use some other ways to style the elements.

I prefer markdown's (and markdoc's) approach to styling, in that it doesn't really try to define any. My take on this is that stylings of the output should be be coupled with the toolings, as opposed to being bolted into the language standards.

edit: formatting

saltymimir

4 anni fadiscuss

Our experience with this has been the complete opposite. Apple's reliance on real human reviewers was precisely the issue for us. Your mileage may vary by a lot depending on the region you operate in.

Google's process, while not perfect, made it at least transparent that the process is going to take a while. And at least in our case, the process involved some real people at the end to point out the specific concerns that actually lead to actionables.

Apple's process is slow, and seems to be worse in my region; it's clear that they're severely understaffed here. We spent a good 30 days trying to fix a crash, and they were all spent for either submitting a new app version to Apple, waiting for the review to finish, getting it rejected for some very vague reasons, and waiting for a response from them to elaborate their reasons for the rejection.

saltymimir

4 anni fadiscuss

I had the opposite experience with Syncthing + Logseq. The way Logseq is designed, refreshing the entire graph is a must before you even consider editing your pages in another device. I forgot to do it once and after some time editing my pages, I realized that I lost a good chunk of my notes from several days back.

saltymimir

4 anni fadiscuss

They do have an Android app now, though? You can download the apk from their github releases page[1], if you don't mind sideloading apps.

[1]: https://github.com/logseq/logseq/releases

saltymimir

4 anni fadiscuss

Congrats! Been using the insider builds, so I've used the overhauled UI for quite a while. What I've been waiting for is the mobile app, and as expected you folks delivered. Hands down the best note-taking experience in mobile platforms, and it's not even close.

I'm still using Logseq as my PKMS app by default, but seeing how active you've been with pushing genuinely new (and useful) changes, I might just reconsider this.

saltymimir

4 anni fadiscuss

If you don't mind the extra setup, you can block domains at DNS level. Something like Pi-hole[1] or NextDNS[2] will prevent those tracker requests in all apps in your device and/or network.

1: https://pi-hole.net

2: https://nextdns.io

saltymimir

4 anni fadiscuss

Save for the fact that it doesn't support non-Apple platforms, NNW is pretty much perfect for me. No upsells of extra features or even a need to create an account. Just a very good, clean, and simple reader.

saltymimir

4 anni fadiscuss

Seconding this, and if anyone is willing to add a performance angle in their assessment, that'd be great too.

I haven't evaluated Markdoc fully, but I did have some experience using both MDX 2 and markdown-it—the parser used by Markdoc. I can honestly say that MDX 2 (and remark for that matter) is a lot slower in parsing bigger markdown files. Anecdotally, I saw markdown-it outperforming remark by a factor of 20 in terms of generating the file's AST alone.

Markdoc with its added complexity may add some penalty to the parsing performance, but does it do so to a point where it gets considerably closer to MDX? I doubt that this is the case, but would love to see someone verify this.

saltymimir

4 anni fadiscuss

I maintain something similar today, and I'm guessing that the OP uses some combination of the following libraries too (?):

- Readability (https://github.com/mozilla/readability) to strip down the page's HTML to a bare minimum.

- Turndown.js (https://github.com/mixmark-io/turndown) to convert the plain HTML to a markdown format with the GFM plugins enabled.

- Puppeteer (https://github.com/puppeteer/puppeteer) to download the page.

It costs me only several cents to parse an entire page, and I think OP can make some money out of this if they get the pricing right.

Also, some unsolicited feedbacks on the API:

- An option to enable/disable javascript would be great, since not all pages actually need to have it enabled to be parsable.

- You can probably tweak the header of the headless browser to bypass the paywalls of some sites. Some are as simple as setting the useragent to a crawler bot (like `googlebot`).

- Maybe an option to fill in the front matter (https://jekyllrb.com/docs/front-matter/) with a metadata given in the payload?

saltymimir

Submissions

Rich Harris on frameworks, the web, and the edge (2023) [video]

Glamorous Toolkit: A Moldable Development Environment

Language Evolution: Problems, and What Can We Do About It?

Anders Hejlsberg and Lars Bak: TypeScript, JavaScript, and Dart (2013) [video]

Moldable Development

RAADS–R: Self-report questionnaire to identify adult autistics (2020)

Ktistec: Single user ActivityPub server backed by SQLite

Preview.js: Preview UI Components in IDEs

Incident with Actions and Codespaces

Succinct Data Structures: Cramming 80k words into a JavaScript file (2011)

comments